Google Flights Scraper

Scrape live Google Flights fares as clean JSON, CSV or Excel. Each row has the price, airline, departure and arrival times, stop count, total duration and every layover.

Give it a list of origins, destinations and dates. Every combination is searched, so one run can build a whole fare calendar, such as LHR to JFK every day in September in business class.

One way, round trip and multi-city. Adults, children and infants. Alliance and stop filters. Over a hundred currencies.

Quick start

Press Start. The form is already filled in with a real search, LHR to JFK thirty days out, so the first run works without you typing anything. It returns roughly 25 to 30 itineraries in about ten seconds and costs around $0.03.

Then change three fields to your own search:

Field	Example
From	LHR, or several codes to compare departure points
To	JFK, or several
Departure dates	+30, or 2026-11-10

To see the cost before you spend anything, tick Dry run. It reports how many results you would get and what they would cost, and it never charges you.

The status line is the estimate. A dry run writes no rows, so the results table stays empty, and the small figure beside the status is Apify's platform usage, which the developer pays under pay-per-event pricing, not you.

Dates that do not go stale

Every date field takes a calendar date such as 2026-11-10, or a number of days from today such as +30.

Use the relative form on a schedule. A daily run pinned to a calendar date returns fares until that date arrives and then returns nothing, with no error to tell you why. +30 still means thirty days out a year later, so a scheduled run needs no maintenance.

// A rolling fare tracker: the next five departure days, re-checked every day.
{ "origins": ["LHR"], "destinations": ["JFK"], "departureDates": ["+1", "+2", "+3", "+4", "+5"] }

What you get

One row per itinerary.

Field	Description
price	Lowest total fare for the whole party, in your chosen currency
currency	ISO code, such as USD, GBP or EUR
tripType	one-way, round-trip or multi-city
priceBasis	How to read price: one-way, whole-trip or per-leg
legIndex / legCount	Which leg this row prices, out of how many
airline / airlines	Marketing airline, and every airline on the itinerary
operatedBy	The carrier actually flying it, when it differs from the one selling
stops	0 for nonstop
departureAirport / departureTime	Airport name and local 24-hour time
departureDate / arrivalDate	Full ISO dates, so an overnight arrival is unambiguous
arrivalAirport / arrivalTime	Airport name and local 24-hour time
durationMinutes	Total journey time
overnight	true when it lands on a later day than it left
layovers[]	Each connection's airport, city and wait in minutes
tripLegs[]	Every leg that was priced, so a round-trip row explains itself
returnDate	The return that was searched, or null on a one-way
searchUrl	The exact Google Flights search. Paste it into a browser and check
origin / destination / searchDate	What was searched, so rows from a multi-route run stay self-describing
adults / children / infantsInSeat / infantsOnLap	The party this fare was quoted for

Example output

These are real rows, not a mock-up. LHR to JFK and LHR to BOS, round trip, economy, out on 10 November and back on 17 November 2026, scraped on 22 September 2026. The run returned 36 itineraries across 9 airlines, 28 of them nonstop, priced from $479 to $673, in 13 seconds.

Price	Route	Airline	Stops	Departs	Arrives	Duration
$479	LHR to BOS	Icelandair	1 stop	12:15	17:50	34h 35m
$494	LHR to JFK	SWISS	1 stop	06:00	13:15	12h 15m
$498	LHR to JFK	Austrian	1 stop	06:00	14:15	13h 15m
$537	LHR to JFK	JetBlue	Nonstop	07:25	10:32	8h 07m
$537	LHR to JFK	British Airways	Nonstop	08:20	11:20	8h 00m
$537	LHR to JFK	Virgin Atlantic	Nonstop	09:05	12:25	8h 20m

One row in full, exactly as it is written to your dataset, with nothing trimmed:

{
  "searchUrl": "https://www.google.com/travel/flights/search?tfs=CBwQAhoeEgoyMDI2LTExLTEwagcIARIDTEhScgcIARIDSkZLGh4SCjIwMjYtMTEtMTdqBwgBEgNKRktyBwgBEgNMSFJAAUgBcAGCAQsI____________AZgBAQ&hl=en&gl=US&curr=USD",
  "tripType": "round-trip",
  "origin": "LHR",
  "destination": "JFK",
  "searchDate": "2026-11-10",
  "returnDate": "2026-11-17",
  "tripLegs": [
    { "origin": "LHR", "destination": "JFK", "date": "2026-11-10" },
    { "origin": "JFK", "destination": "LHR", "date": "2026-11-17" }
  ],
  "legIndex": 1,
  "legCount": 1,
  "priceBasis": "whole-trip",
  "cabinClass": "economy",
  "adults": 1,
  "children": 0,
  "infantsInSeat": 0,
  "infantsOnLap": 0,
  "price": 494,
  "currency": "USD",
  "airline": "SWISS",
  "airlines": ["SWISS"],
  "operatedBy": null,
  "stops": 1,
  "departureAirport": "Heathrow Airport",
  "departureTime": "06:00",
  "departureDate": "2026-11-10",
  "arrivalAirport": "John F. Kennedy International Airport",
  "arrivalTime": "13:15",
  "arrivalDate": "2026-11-10",
  "durationMinutes": 735,
  "overnight": false,
  "layovers": [{ "airport": "Zurich Airport", "city": "Zürich", "durationMinutes": 75 }],
  "id": "round-trip|1|LHR|JFK|2026-11-10|2026-11-17|economy|1|0|0|0|2026-11-10|06:00|2026-11-10|13:15|1|SWISS",
  "actorName": "google-flights",
  "scrapedAt": "2026-09-22T15:20:45.248Z"
}

Paste a row's searchUrl into a browser and you land on the search it came from. That is how you check any fare here against the source.

Input

Setting	What it does
Trip type	One way, round trip or multi-city.
From / To	Pick from 1,149 major airports and 22 whole cities, or type any IATA code.
Departure dates	YYYY-MM-DD, or +30 for days from today. Add a range to find the cheapest day.
Return dates	Round trip only, same formats. Every departure is paired with every return.
Trip length in nights	Easier than listing returns: enter 7 and 14 to price both from every departure.
Multi-city legs	Multi-city only. One row per leg, each with a from, a to and a date.
Passengers	Adults, children aged 2 to 11, infants in seat, infants on lap. Up to 9 seated.
Cabin	Economy, premium economy, business or first.
Maximum stops	Any, nonstop only, 1 stop or fewer, 2 stops or fewer. Applied by Google itself.
Currency	Over a hundred, each verified end to end against the live page.
Dry run	Estimates the result count and cost, then stops. Costs nothing.
Maximum results	Hard cap on rows. The run stops the moment it is reached.
Maximum spend (USD)	Hard cap on cost. The run stops before exceeding it.

A city in From or To, such as London, New York, Paris, Tokyo and eighteen more, searches every airport in it at once. LON to NYC returns Heathrow, Gatwick and Stansted to JFK, Newark and LaGuardia in one run, which is usually the search you actually want. Any code not in the list can still be typed.

There is no limit on how many origins, destinations or dates you provide. Cost is bounded by Maximum results and Maximum spend, and Dry run tells you the real number for free before you commit.

A single route on a single date returns roughly 25 to 30 itineraries, which is a useful number to size Maximum results against.

How prices are quoted

Every row carries a priceBasis field, so you never have to work out what a price covers by checking which trip type produced it.

Trip type	priceBasis	What price means
One way	one-way	That flight
Round trip	whole-trip	The entire return journey, all passengers
Multi-city	per-leg	That leg alone, priced as a one-way

On a round trip, Google prices each outbound option at the whole-trip total, because you pick an outbound before it will show you returns. A round-trip row therefore describes the outbound flight and prices the whole journey.

On multi-city, each leg is searched and priced separately, so you get every option for every leg. That is more than Google's own multi-city view gives you, which lists first-leg choices only. A three-city trip returned 44 bookable itineraries in testing where Google's multi-city page showed 26 first-leg options.

The caveat, and the reason priceBasis exists: the sum of the legs is not a single multi-city ticket, and a combined fare is often cheaper. If you need one ticket, use these prices to find the right flights and then price that itinerary directly. Adding three one-ways together and calling it a multi-city fare would be the wrong number.

Accuracy

Fares are read from the itinerary descriptions Google writes into the page for screen-reader users. Those are sentences with named fields rather than an internal feed of unlabelled positional arrays.

That is the difference that matters. If a positional feed shifts by one element, every row still parses, the row count does not change, and the price column starts reporting a number that is not the price. A reworded sentence cannot fail that way. The parse rate collapses, this Actor's health checks fail, the run stops, and you are charged nothing.

Currency is read the same way. The Actor takes the currency from the page rather than trusting the one it asked for, because Google does not reject a currency it cannot serve. It prices in a different one instead. When that happens the run stops and names the currency it got, rather than billing you for prices in the wrong money.

Checked against the leading alternative in this category, this Actor found a real mispricing in it: two Zurich-connecting SWISS itineraries an hour apart, with the overnight departure's cheaper fare attached to the morning flight. That is the failure described above, in production. This Actor also returns layover airports and durations, operating carriers, and correct next-day arrival dates, which that Actor does not.

Limitations

Multi-city is priced leg by leg, not as one ticket. This is deliberate. Google's own multi-city cannot be read reliably without a browser. Asked one way it returns an empty page. Asked another it returns a healthy-looking page that has priced a plain round trip instead, at prices identical to the pound, for a search that never happened. Each leg is therefore searched on its own. If another scraper offers combined multi-city fares at HTTP speed, compare one of its open-jaw prices against the equivalent round trip before trusting it.

Nine seated passengers is the maximum. Above that Google returns one-passenger fares without saying so, so the search is refused rather than answered wrongly.

A route with no flights fails the run instead of returning an empty dataset. The page this Actor reads never says "no flights". A route with no service renders the same endless loading shell as a rate-limited request, so an empty answer cannot be told apart from a blocked one. Reporting "no flights exist" on evidence that might mean "we were throttled" would be a guess, so the run fails, says which it cannot rule out, and charges you nothing. If you think the route should have flights, run it again.

No flight numbers. Google does not put them in the data this Actor reads.

English only. Results are requested in English so the fields can be read reliably.

Pricing

Pay per result, and nothing else. No per-run fee, no charge for compute, no charge for retries.

Plan	Per 1,000 results
Free	$1.00
Bronze	$0.90
Silver	$0.80
Gold	$0.70
Platinum	$0.70
Diamond	$0.70

You pay only for results you actually receive. If the run fails, returns nothing, or Google changes the page and the health checks catch it, you are charged nothing. Duplicate itineraries are removed before billing, so the same flight is never sold to you twice.

Tick Dry run first to see the cost before you spend anything.

Support and feature requests

Email actors@montyburrows.com if you find a bug, need an extra field, or want a filter that is not here yet. Feature requests are welcome and usually quick.